Abstract
OBJECTIVE: We examined if thyroid autoimmunity is relevant to the relationship between maternal thyroid stimulating hormone (TSH) levels and pregnancy outcomes. DESIGN: Retrospective cohort analysis of data from 2 randomized controlled trials (RCTs). SUBJECTS: Participants of the Pregnancy in Polycystic Ovary Syndrome (PPCOS II, n = 746) and the Assessment of Multiple Intrauterine Gestations from Ovarian Stimulation (AMIGOS, n = 832 with unexplained infertility) RCTs. EXPOSURE: Pre-RCT intervention levels of TSH at threshold of ≥2.0 mU/L and thyroid peroxidase antibody (TPO-Ab) at titer threshold of ≥30 U/mL. MAIN OUTCOME MEASURES: Live birth (primary outcome), pregnancy loss, and preterm birth (secondary outcomes). Generalized linear model (GLM) analyses examined the relationship between exposure to TSH and TPO-Ab at specified thresholds with the specified outcomes; covariates adjusted for included age, body mass index, race, ethnicity, education, smoking, duration of infertility, PCOS (vs. unexplained infertility), and randomized intervention arm in the respective RCTs. RESULTS: On adjusted analyses, live birth was significantly reduced in the exposed population (those with TSH ≥2.0 mU/L and TPO-Ab ≥30 U/mL, n = 117/1,578, 7.4%, adjusted risk ratio [ARR]: 0.55; 95% CI: 0.35-0.87) compared with the unexposed (those with TSH <2.0 mU/L and TPO-Ab <30 U/mL, n = 865/1,578, 54.8%). Furthermore, the risk of pregnancy loss and of early preterm birth (<32 weeks) was significantly higher in the exposed compared with the unexposed (ARR for pregnancy loss was 1.66; 95% CI: 1.14-2.42, and ARR for early preterm birth was 4.82, 95% CI: 1.53-15.19). CONCLUSION: In women with TPO-Ab titers ≥30 U/mL, pregnancy outcomes may be compromised at TSH threshold of ≥2 mU/L. These findings of an interaction between TSH and TPO for pregnancy outcomes merit further investigation in prospective studies. TRIAL REGISTRATION NUMBER: NCT00719186 and NCT01044862.